Method for detecting grammatical errors, error detection device for same and computer-readable recording medium having method recorded thereon

Abstract

An apparatus for detecting grammatical errors includes: a sentence analyzer to break up an input sentence into units of morphemes; an example builder to break up example text into units of morphemes and build an example-based index database (DB); and an error detector to generate morpheme sequences by binding the morphemes broken up from the input sentence in a preset window (n-window) size, the generated morpheme sequences comprising forward morpheme sequences and backward morpheme sequences, determine frequencies of appearance of morpheme sequences identical to the forward morpheme sequences and backward morpheme sequences by searching the example-based index DB, and detect grammatical errors in the input sentence by combining the determined frequencies of appearance.

What is claimed is: 
     
       1. A non-transitory computer-readable recording medium storing a program for executing a method for detecting grammatical errors, the program, when executed by a processor included in an apparatus for detecting grammatical errors, causing the processor to execute:
 breaking up an input sentence into units of morphemes; 
 generating morpheme sequences by binding the broken-up morphemes in a preset window size set by the processor, the generated morpheme sequences comprising forward morpheme sequences and backward morpheme sequences; 
 determining frequencies of appearance of morpheme sequences identical to forward morpheme sequences and backward morpheme sequences of the generated morpheme sequences by searching an example-based index database (DB); and 
 detecting grammatical errors in the input sentence by combining the determined frequencies of appearance. 
 
     
     
       2. The non-transitory computer-readable recording medium of  claim 1 , wherein the generating of the morpheme sequences comprises
 generating (i) the forward morpheme sequences by binding morphemes arranged in a forward direction from the respective morphemes broken up from the input sentence in the preset window size, and (ii) the backward morpheme sequences by binding morphemes arranged in a backward direction from the respective morphemes broken up from the input sentence in the preset window size. 
 
     
     
       3. The non-transitory computer-readable recording medium of  claim 2 , wherein the detecting of the errors comprises:
 searching whether or not the generated morpheme sequences are in the example-based index DB; 
 extracting the frequencies of appearance at which the forward morpheme sequences and the backward morpheme sequences are identical to morpheme sequences retrieved from the example-based index DB, and calculating suitability scores based on the extracted frequencies of appearance; and 
 detecting grammatical errors by determining whether the calculated suitability scores are smaller than a threshold or whether the calculated suitability scores are smaller than the threshold by a predetermined value or more. 
 
     
     
       4. The non-transitory computer-readable recording medium of  claim 3 , wherein the calculating of the suitability scores comprises
 calculating respective suitability scores based on the frequencies of appearance of the forward morpheme sequences and the backward morpheme sequences, and 
 calculating the suitability scores in an integrated manner by considering the respective suitability scores. 
 
     
     
       5. The non-transitory computer-readable recording medium of  claim 3 , wherein the calculating of the suitability scores comprises
 calculating the suitability scores by applying different weights according to frequencies of appearance at which the respective morphemes are identically arranged in the morpheme sequences retrieved from the example-based index DB. 
 
     
     
       6. The non-transitory computer-readable recording medium of  claim 1 , wherein the detecting of the errors comprises
 when frequencies of appearance of the forward morpheme sequences are different from frequencies of appearance of the backward morpheme sequences, detecting the errors based on higher frequencies of appearance among frequencies of appearance of the forward morpheme sequences and the backward morpheme sequences. 
 
     
     
       7. The non-transitory computer-readable recording medium of  claim 6 , wherein the detecting of the grammatical errors based on the higher frequencies of appearance comprises:
 calculating suitability scores using the higher frequencies of appearance; and 
 detecting the grammatical errors by determining whether the suitability scores are smaller than a threshold or whether the suitability scores are smaller than the threshold by a predetermined value or more. 
 
     
     
       8. An apparatus for detecting grammatical errors, the apparatus comprising:
 a sentence analyzer, implemented by a processor s and/or application-specific integrated circuits (ASICs), configured to break up an input sentence into units of morphemes; 
 an example builder, implemented by one or more processors and/or application-specific integrated circuits (ASICs), configured to break up example text into units of morphemes and build an example-based index database (DB); and 
 an error detector, implemented by one or more processors and/or application-specific integrated circuits (ASICs), configured to
 generate morpheme sequences by binding the morphemes broken up from the input sentence in a preset window (n-window) size, the generated morpheme sequences comprising forward morpheme sequences and backward morpheme sequences, 
 determine frequencies of appearance of morpheme sequences identical to the forward morpheme sequences and backward morpheme sequences by searching the example-based index DB, and 
 detect grammatical errors in the input sentence by combining the determined frequencies of appearance. 
 
 
     
     
       9. The apparatus of  claim 8 , wherein the error detector comprises:
 an n-gram generation unit configured to generate the forward and backward morpheme sequences by binding morphemes arranged in a forward direction and a backward direction, respectively, from the respective morphemes broken up from the input sentence in the preset window size; 
 a search unit configured to search whether the forward morpheme sequences and the backward morpheme sequences generated by the n-gram generation unit are in the example-based index DB; 
 a suitability calculation unit configured to extract the frequencies of appearance at which the forward morpheme sequences and the backward morpheme sequences generated by the n-gram generation unit are identical to morpheme sequences retrieved by the search unit, and calculate suitability scores based on the extracted frequencies of appearance; and 
 an error determination unit configured to detect grammatical errors by determining whether the suitability scores calculated by the suitability calculation unit are smaller than a threshold or whether the suitability scores are smaller than the threshold by a predetermined value or more. 
 
     
     
       10. The apparatus of  claim 9 , wherein the suitability calculation unit is configured to
 calculate respective suitability scores of the forward morpheme sequences and the backward morpheme sequences generated by the n-gram generation unit, and 
 calculate the suitability scores in an integrated manner by considering the calculated respective suitability scores. 
 
     
     
       11. The apparatus of  claim 9 , wherein the suitability calculation unit is configured to calculate the suitability scores by applying different weights according to frequencies of appearance at which respective morphemes are identically arranged in the morpheme sequences retrieved by the search unit. 
     
     
       12. The apparatus of  claim 8 , wherein, when frequencies of appearance of the forward morpheme sequences are different from frequencies of appearance of the backward morpheme sequences, the error detector is configured to detect the grammatical errors in the sentence based on higher frequencies of appearance among the frequencies of appearance of the forward morpheme sequences and the frequencies of the backward morpheme sequences. 
     
     
       13. The apparatus of  claim 12 , wherein the error detector is configured to calculate suitability scores using the higher frequencies of appearance, and detects grammatical errors by determining whether the suitability scores are smaller than a threshold or whether the suitability scores are smaller than the threshold by a predetermined value or more. 
     
     
       14. A method performed by an apparatus for detecting grammatical errors, the apparatus including a processor to execute the method comprising:
 breaking up an input sentence into units of morphemes; 
 generating morpheme sequences by binding the broken-up morphemes in an preset window size set by the processor, the generated morpheme sequences comprising forward morpheme sequences and backward morpheme sequences; 
 determining frequencies of appearance of morpheme sequences identical to the forward morpheme sequences and backward morpheme sequences of the generated morpheme sequences by searching an example-based index database (DB); and 
 detecting grammatical errors in the input sentence by combining the determined frequencies of appearance. 
 
     
     
       15. The method of  claim 14 , wherein the generating of the morpheme sequences comprises
 generating (i) the forward morpheme sequences by binding morphemes arranged in a forward direction from the respective morphemes broken up from the input sentence in the preset window size, and (ii) the backward morpheme sequences by binding morphemes arranged in a backward direction from the respective morphemes broken up from the input sentence in the preset window size. 
 
     
     
       16. The method of  claim 15 , wherein the detecting of the errors comprises:
 searching whether or not the generated morpheme sequences are in the example-based index DB; 
 extracting the frequencies of appearance at which the forward morpheme sequences and the backward morpheme sequences are identical to morpheme sequences retrieved from the example-based index DB, and calculating suitability scores based on the extracted frequencies of appearance; and 
 detecting grammatical errors by determining whether the calculated suitability scores are smaller than a threshold or whether the calculated suitability scores are smaller than the threshold by a predetermined value or more. 
 
     
     
       17. The method of  claim 16 , wherein the calculating of the suitability scores comprises
 calculating respective suitability scores based on the frequencies of appearance of the forward morpheme sequences and the backward morpheme sequences, and 
 calculating the suitability scores in an integrated manner by considering the calculated respective suitability scores. 
 
     
     
       18. The method of  claim 16 , wherein the calculating of the suitability scores comprises
 calculating the suitability scores by applying different weights according to frequencies of appearance at which the respective morphemes are identically arranged in the morpheme sequences retrieved from the example-based index DB. 
 
     
     
       19. The method of  claim 14 , wherein the detecting of the errors comprises
 when frequencies of appearance of the forward morpheme sequences are different from frequencies of appearance of the backward morpheme sequences, detecting the errors based on higher frequencies of appearance among frequencies of appearance of the forward morpheme sequences and the backward morpheme sequences. 
 
     
     
       20. The method of  claim 19 , wherein the detecting of the grammatical errors based on the higher frequencies of appearance comprises:
 calculating suitability scores using the higher frequencies of appearance; and 
 detecting the grammatical errors by determining whether the suitability scores are smaller than a threshold or whether the suitability scores are smaller than the threshold by a predetermined value or more.